INDEX

Explanations

project

This neuron activates on numerical tokens and measures—i.e. digit sequences and size/quantity figures in the text.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

śni

-0.79

λους

-0.75

닷

-0.75

Handlung

-0.75

܀

-0.74

tapeta

-0.73

负担

-0.73

attore

-0.73

umenten

-0.73

艏

-0.72

POSITIVE LOGITS

 project

7.75

 projects

6.34

project

5.16

 proyecto

4.81

项目

4.81

 проекта

4.56

PROJECT

4.44

プロジェクト

4.31

 پروژه

4.25

 Project

4.22

Activations Density 0.197%